Jochen Kunz wrote:
> On Wed, 23 Feb 2005 14:33:42 -0500
> Henry Bent <Henry.Bent@oberlin.edu> wrote:
> 
> 
>>1. the GENERIC kernel config has a commented entry for "siop* at gsc?"
>>but this driver doesn't work at all.
> 
> This is well possible. There are not that many supported machines with
> this kind of SCSI adapter. I have none. Without hardware it is hard to
> debug things...
> 
> Keep in mind that this FW SCSI bus is most likely High Voltage
> Differential SCSI. So you can't connect Single Ended or Low Voltage
> Differential devices to it.

Indeed, the drive sled I have is marked "9.1 DF."  Fortunately I had an 
HVD drive to put in it.


>>2. I have a J3515 10/100 ethernet card that doesn't appear to be 
>>supported at all.  The chip is a DEC 21140-AE.  Is there any chance of
>>a driver for this card?  It would make the system a LOT more usable to
>>me!
> 
> This card has a Dino GSC to PCI bridge and the mentioned Tulip chip on
> it. Support may be trivial to fix, but I would need access to this
> hardware. Dino has a special "Card Mode" that may cause some
> complications, but it should be doable. I tried to get one of those
> cards on ePay only to write a driver for it, but I got outbid and I
> did't want to sped lots of money for a card that I have no real need
> for.
> 
> You can try a hack:
> Replace HPPA_TYPE_BRIDGE with HPPA_TYPE_ADMA and HPPA_BRIDGE_DINO with
> HPPA_ADMA_HSCPCI in src/sys/arch/hp700/dev/dino.c, recompile a kernel
> and boot it. 

Here's what I got with that change:

...
dino0 at phantomas0 hpa 0xfff88000 path 8/8 irq 28 ipl 1: Mojo V4.0
pci0 at dino0 bus 0
FLT CBFB

That last line was actually overlaid on the "pci0 at dino0" line.  At 
that point the machine was wedged, and I had to break out with ctrl-B. 
Looks somewhat promising though, given that it identified the dino 
bridge correctly.

-Henry Bent
henry.bent@oberlin.edu